当前位置: 首页 > 知识库问答 >
问题:

有人能给我举个例子,说明如何使用PDFBox提取“单词”的坐标吗

邓越泽
2023-03-14

有人能给我一个如何用PDFBox提取“单词”坐标的例子吗

我正在使用此链接提取单个角色的位置:https://www.tutorialkart.com/pdfbox/how-to-extract-coordinates-or-position-of-characters-in-pdf/

我正在使用此链接提取单词:https://www.tutorialkart.com/pdfbox/extract-words-from-pdf-document/

我被困在获取整个单词的坐标上。

共有2个答案

劳彦
2023-03-14

您可以创建CustomPDFTextStripper,它扩展了PDFTextStripper并覆盖受保护的void writeString(字符串文本,列表

下面的完整示例还包含生成的边界框的绘图。

package com.example;

import lombok.Value;
import org.apache.pdfbox.pdmodel.PDDocument;
import org.apache.pdfbox.rendering.PDFRenderer;
import org.apache.pdfbox.text.PDFTextStripper;
import org.apache.pdfbox.text.TextPosition;
import org.junit.Ignore;
import org.junit.Test;

import javax.imageio.ImageIO;
import java.awt.*;
import java.awt.image.BufferedImage;
import java.io.*;
import java.util.ArrayList;
import java.util.List;
import java.util.stream.Collectors;

public class PdfBoxTest {

    private static final String BASE_DIR_PATH = "C:\\Users\\Milan\\50330484";
    private static final String INPUT_FILE_PATH = "input.pdf";
    private static final String OUTPUT_IMAGE_PATH = "output.jpg";
    private static final String OUTPUT_BBOX_IMAGE_PATH = "output-bbox.jpg";

    private static final float FROM_72_TO_300_DPI = 300.0f / 72.0f;

    @Test
    public void run() throws Exception {
        pdfToImage();
        drawBoundingBoxes();
    }

    @Ignore
    @Test
    public void pdfToImage() throws IOException {
        PDDocument document = PDDocument.load(new File(BASE_DIR_PATH, INPUT_FILE_PATH));
        PDFRenderer renderer = new PDFRenderer(document);
        BufferedImage image = renderer.renderImageWithDPI(0, 300);
        ImageIO.write(image, "JPEG", new File(BASE_DIR_PATH, OUTPUT_IMAGE_PATH));
    }

    @Ignore
    @Test
    public void drawBoundingBoxes() throws IOException {

        PDDocument document = PDDocument.load(new File(BASE_DIR_PATH, INPUT_FILE_PATH));

        List<WordWithBBox> words = getWords(document);

        draw(words);
    }

    private List<WordWithBBox> getWords(PDDocument document) throws IOException {

        CustomPDFTextStripper customPDFTextStripper = new CustomPDFTextStripper();
        customPDFTextStripper.setSortByPosition(true);
        customPDFTextStripper.setStartPage(0);
        customPDFTextStripper.setEndPage(1);

        Writer writer = new OutputStreamWriter(new ByteArrayOutputStream());
        customPDFTextStripper.writeText(document, writer);

        List<WordWithBBox> words = customPDFTextStripper.getWords();

        return words;
    }

    private void draw(List<WordWithBBox> words) throws IOException {

        BufferedImage bufferedImage = ImageIO.read(new File(BASE_DIR_PATH, OUTPUT_IMAGE_PATH));

        Graphics2D graphics = bufferedImage.createGraphics();

        graphics.setColor(Color.GREEN);

        List<Rectangle> rectangles = words.stream()
                .map(word -> new Rectangle(word.getX(), word.getY(), word.getWidth(), word.getHeight()))
                .collect(Collectors.toList());
        rectangles.forEach(graphics::draw);

        graphics.dispose();

        ImageIO.write(bufferedImage, "JPEG", new File(BASE_DIR_PATH, OUTPUT_BBOX_IMAGE_PATH));
    }

    private class CustomPDFTextStripper extends PDFTextStripper {

        private final List<WordWithBBox> words;

        public CustomPDFTextStripper() throws IOException {
            this.words = new ArrayList<>();
        }

        public List<WordWithBBox> getWords() {
            return new ArrayList<>(words);
        }

        @Override
        protected void writeString(String text, List<TextPosition> textPositions) throws IOException {

            String wordSeparator = getWordSeparator();
            List<TextPosition> wordTextPositions = new ArrayList<>();

            for (TextPosition textPosition : textPositions) {
                String str = textPosition.getUnicode();
                if (wordSeparator.equals(str)) {
                    if (!wordTextPositions.isEmpty()) {
                        this.words.add(createWord(wordTextPositions));
                        wordTextPositions.clear();
                    }
                } else {
                    wordTextPositions.add(textPosition);
                }
            }

            super.writeString(text, textPositions);
        }

        private WordWithBBox createWord(List<TextPosition> wordTextPositions) {

            String word = wordTextPositions.stream()
                    .map(TextPosition::getUnicode)
                    .collect(Collectors.joining());

            int minX = Integer.MAX_VALUE;
            int minY = Integer.MAX_VALUE;
            int maxX = Integer.MIN_VALUE;
            int maxY = Integer.MIN_VALUE;

            for (TextPosition wordTextPosition : wordTextPositions) {

                minX = Math.min(minX, from72To300Dpi(wordTextPosition.getXDirAdj()));
                minY = Math.min(minY, from72To300Dpi(wordTextPosition.getYDirAdj() - wordTextPosition.getHeightDir()));
                maxX = Math.max(maxX, from72To300Dpi(wordTextPosition.getXDirAdj() + wordTextPosition.getWidthDirAdj()));
                maxY = Math.max(maxY, from72To300Dpi(wordTextPosition.getYDirAdj()));
            }

            return new WordWithBBox(word, minX, minY, maxX - minX, maxY - minY);
        }
    }

    private int from72To300Dpi(float f) {
        return Math.round(f * FROM_72_TO_300_DPI);
    }

    @Value
    private class WordWithBBox {
        private final String word;
        private final int x;
        private final int y;
        private final int width;
        private final int height;
    }
}

注:

如果您对其他选项感兴趣,也可以查看Popple

PDF转图像

pdftoppm -r 300 -jpeg input.pdf output

生成一个XHTML文件,其中包含文件中每个单词的边界框信息。

pdftotext -r 300 -bbox input.pdf

拓拔阎宝
2023-03-14

通过收集所有构建单词的TextPosition对象并组合它们的边界框,可以提取单词的坐标。

按照您参考的两个教程的思路实现这一点,您可以像这样扩展PDFTextStripper

public class GetWordLocationAndSize extends PDFTextStripper {
    public GetWordLocationAndSize() throws IOException {
    }

    @Override
    protected void writeString(String string, List<TextPosition> textPositions) throws IOException {
        String wordSeparator = getWordSeparator();
        List<TextPosition> word = new ArrayList<>();
        for (TextPosition text : textPositions) {
            String thisChar = text.getUnicode();
            if (thisChar != null) {
                if (thisChar.length() >= 1) {
                    if (!thisChar.equals(wordSeparator)) {
                        word.add(text);
                    } else if (!word.isEmpty()) {
                        printWord(word);
                        word.clear();
                    }
                }
            }
        }
        if (!word.isEmpty()) {
            printWord(word);
            word.clear();
        }
    }

    void printWord(List<TextPosition> word) {
        Rectangle2D boundingBox = null;
        StringBuilder builder = new StringBuilder();
        for (TextPosition text : word) {
            Rectangle2D box = new Rectangle2D.Float(text.getXDirAdj(), text.getYDirAdj(), text.getWidthDirAdj(), text.getHeightDir());
            if (boundingBox == null)
                boundingBox = box;
            else
                boundingBox.add(box);
            builder.append(text.getUnicode());
        }
        System.out.println(builder.toString() + " [(X=" + boundingBox.getX() + ",Y=" + boundingBox.getY()
                 + ") height=" + boundingBox.getHeight() + " width=" + boundingBox.getWidth() + "]");
    }
}

(内部类)

然后像这样运行:

PDDocument document = PDDocument.load(resource);
PDFTextStripper stripper = new GetWordLocationAndSize();
stripper.setSortByPosition( true );
stripper.setStartPage( 0 );
stripper.setEndPage( document.getNumberOfPages() );

Writer dummy = new OutputStreamWriter(new ByteArrayOutputStream());
stripper.writeText(document, dummy);

(ExtractWords坐标测试TestExtractWordsForgoodju

应用于apache。pdf示例教程使用的内容如下:

2017-8-6 [(X=26.004425048828125,Y=22.00372314453125) height=5.833024024963379 width=36.31868362426758]
Welcome [(X=226.44479370117188,Y=22.00372314453125) height=5.833024024963379 width=36.5999755859375]
to [(X=265.5881652832031,Y=22.00372314453125) height=5.833024024963379 width=8.032623291015625]
The [(X=276.1641845703125,Y=22.00372314453125) height=5.833024024963379 width=14.881439208984375]
Apache [(X=293.5890197753906,Y=22.00372314453125) height=5.833024024963379 width=29.848846435546875]
Software [(X=325.98126220703125,Y=22.00372314453125) height=5.833024024963379 width=35.271636962890625]
Foundation! [(X=363.7962951660156,Y=22.00372314453125) height=5.833024024963379 width=47.871429443359375]
Custom [(X=334.0334777832031,Y=157.6195068359375) height=4.546705722808838 width=25.03936767578125]
Search [(X=360.8929138183594,Y=157.6195068359375) height=4.546705722808838 width=22.702728271484375]
 类似资料:
  • 嗨,这个问题是指之前的帖子: 有人能给我举个例子,说明如何使用PDFBox提取“单词”的坐标吗 我使用的是PDFBOX 2.0.10 我已成功编译了组合代码,但在尝试运行示例时出现异常错误。 提供的解决方案没有标准的主方法,这是我感到困惑的地方。 有人能告诉我怎样才能成功运行组合代码吗。 堆栈跟踪 可以在这里找到坐标https://github.com/mkl-public/testarea-pd

  • 本文向大家介绍举例说明如何使用WebSQL?相关面试题,主要包含被问及举例说明如何使用WebSQL?时的应答技巧和注意事项,需要的朋友参考一下

  • 我正在使用PDFbox来提取PDF文档中单词/字符串的坐标,并且到目前为止已经成功地确定了单个字符的位置。 这将生成一系列包含每个字符位置的行,包括空格,如下所示: 其中“P”是字符。我还没有在PDFbox中找到查找单词的函数,而且我对Java还不够熟悉,无法将这些字符准确地连接回单词中进行搜索,即使空格也包括在内。有没有其他人遇到过类似的情况,如果有,你是如何处理的?我真的只需要单词中第一个字符

  • 本文向大家介绍举例说明clear取值有哪些?相关面试题,主要包含被问及举例说明clear取值有哪些?时的应答技巧和注意事项,需要的朋友参考一下 none 默认值。允许浮动元素出现在两侧。 left 在左侧不允许浮动元素。 right 在右侧不允许浮动元素。 both 在左右两侧均不允许浮动元素。 inherit 从父元素继承 clear 属性的值。

  • 本文向大家介绍举例说明常用的cursor取值有哪些?相关面试题,主要包含被问及举例说明常用的cursor取值有哪些?时的应答技巧和注意事项,需要的朋友参考一下 https://www.w3school.com.cn/cssref/pr_class_cursor.asp 把这个背熟,需要的时候才可以逢场作戏,题目让你说说你常用的,其实你在保证正确的情况下多回答几个是没错的。

  • 本文向大家介绍举例说明如何使用纯html怎么实现下拉提示的功能?相关面试题,主要包含被问及举例说明如何使用纯html怎么实现下拉提示的功能?时的应答技巧和注意事项,需要的朋友参考一下 datalist标签